Arvind Narayanan: Anthropic's capability chart doesn't imply an intelligence explosion
random_walker · x · 2026-09-22
Princeton professor Arvind Narayanan pushes back on speculative readings of Anthropic's capability chart, arguing it doesn't show we're anywhere near an intelligence explosion or superintelligence. His core chain: task delegation ≠ task automation ≠ process automation ≠ faster progress ≠ recursive self-improvement ≠ intelligence explosion. Key precedent: a year ago, many feared/hoped that once AI wrote 100% of code, software output would explode and engineers become obsolete. Teams have now largely hit that milestone, but neither assumption held — humans are still needed and shipping velocity rose without exploding.
More from AGI Musings
- Hot take: Inference on static models that don't learn is massive compute waste — Darpinian · 2026-09-22
- Interview: first conscious AI may 'leapfrog' to sophisticated consciousness — eschwitz · 2026-09-22
- Did you really expect your 2016 job to resemble your 2026 job? — tlakomy · 2026-09-22
- AI safety could be the first nail in the coffin of antitrust's consumer welfare standard — panickssery · 2026-09-22
- birchlse calls researchers' AI risk list "deeply unserious" for ignoring the technology itself — birchlse · 2026-09-22
- Peter Diamandis calls it a 'pandemic of fear,' pushing back on AI extinction concerns — PeterDiamandis · 2026-09-22